PMIScore: An Unsupervised Approach to Quantify Dialogue Engagement
About
High dialogue engagement is a crucial indicator of an effective conversation. A reliable measure of engagement could help benchmark large language models, enhance the effectiveness of human-computer interactions, or improve personal communication skills. However, quantifying engagement is challenging, since it is subjective and lacks a "gold standard". This paper proposes PMIScore, an efficient unsupervised approach to quantify dialogue engagement. It uses pointwise mutual information (PMI), which is the probability of generating a response conditioning on the conversation history. Thus, PMIScore offers a clear interpretation of engagement. As directly computing PMI is intractable due to the complexity of dialogues, PMIScore learned it through a dual form of divergence. The algorithm includes generating positive and negative dialogue pairs, extracting embeddings by large language models (LLMs), and training a small neural network using a mutual information loss function. We validated PMIScore on both synthetic and real-world datasets. Our results demonstrate the effectiveness of PMIScore in PMI estimation and the reasonableness of the PMI metric itself.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Mutual Information Estimation | Synthetic Block | PMIScore ρ0.811 | 6 | |
| Mutual Information Estimation | Synthetic Diagonal | PMI ρ Score0.664 | 6 | |
| Mutual Information Estimation | Synthetic Independent | PMI Score (ρ)0.11 | 6 |